Klein 9B
v1
Showcase images :
Distilled model
768x1152
Euler / normal / 4 steps / CFG 1
Lora strength of 0.6-0.9
Qwen v1 :
"Futurama style" with a strenght of 1
Trained on the first season of Futurama (1999).
Flux v2 edit :
Done. I think it's way better now. Follow prompt better and more flexible.
'An image from the animated TV show Futurama of' (with weight around 0.7 - 0.9)
Flux v1:
I may retrain it one day because it's not as a flexible as I would like.
Feel free to post your best images below. ⬇️
Description
FAQ
Comments (4)
Fine
When I tried Flintstones prompts with the base model, it became clear Flux knows the Flintstones but is nerfed to prevent accurate versions of the show characters from being rendered. These tiny LORAs merely attempt to reactivate skills already present in the base model. To go further and better, the LORA will need to be a lot bigger!